One-Dimensional Zigzag Chains of Cs - ACS Publications - American

Li+(C211)Cs- is the second ceside and third alkalide with a one-dimensional (1D) zigzag chain of alkali metal anions. The distance between adjacent Cs...
0 downloads 0 Views 224KB Size
J. Phys. Chem. B 2006, 110, 12293-12301

12293

One-Dimensional Zigzag Chains of Cs-: The Structures and Properties of Li+(Cryptand[2.1.1])Cs- and Cs+(Cryptand[2.2.2])CsAndrew S. Ichimura,*,‡ Rui H. Huang,† Qingshan Xie,† Philip Morganelli,‡ Amy Burns,† and James L. Dye*,† Department of Chemistry, Michigan State UniVersity, East Lansing, Michigan 48824, and Department of Chemistry and Biochemistry, San Francisco State UniVersity, 1600 Holloway AVenue, San Francisco, California 94132 ReceiVed: February 24, 2006; In Final Form: April 29, 2006

The crystal structure and properties of lithium (cryptand[2.1.1]) ceside, Li+(C211)Cs-, are reported. Li+(C211)Cs- is the second ceside and third alkalide with a one-dimensional (1D) zigzag chain of alkali metal anions. The distance between adjacent Cs- anions, 6 Å, is shorter than the sum of the van der Waals radii, 7 Å. Optical, magic angle spinning NMR, two-probe alternating and direct current conductivity, and electron paramagnetic resonance measurements reveal unique physical properties that result from the overlap of adjacent Cs- wave functions in the chain structure. The properties of cesium (cryptand[2.2.2]) ceside, Cs+(C222)Cs-, were also studied to compare the effects of the subtle geometric changes between the two 1D zigzag chain structures. Li+(C211)Cs- and Cs+(C222)Cs- are both low-band-gap semiconductors with anisotropic reflectivities and large paramagnetic 133Cs NMR chemical shifts relative to Cs-(g). An electronic structure model consistent with the experimental data has sp2-hybridized Cs- within the chain and sp-hybridized chain ends. Ab initio multiconfiguration self-consistent field calculations on the ceside trimer, Cs33-, support this model and indicate a net bonding interaction between nearest neighbors. The buildup of electron density between adjacent Cs- anions is visualized through an electron density difference map constructed by subtracting the density of three cesium atoms from the short Cs33- fragment.

Introduction

CHART 1

Alkalides are a unique class of ionic salts that have alkali metal anions (where M- is Na-, K-, Rb-, or Cs-) as the negative ions and alkali cations embedded in cryptands, crown ethers, or their aza analogues as the positive counterions.1-8 Coulombic attraction between the complexed cations and Manions stabilizes the salt, but we would expect Coulomb repulsion between the anions to keep them separated from one another. Yet, of the 37 published alkalide crystal structures, five have alkali metal anions in close proximity at distances smaller than the sum of the van der Waals radii.9 Two salts, K+(C222)K(where C222 is cryptand[2.2.2], Chart 1) and Rb+(C222)Rbhave very similar structures with pairs of anions surrounded by complexed cations. In addition, the structure of the first sodide dimer, Na22-, was recently reported.10 Even more remarkable are the salts Rb+(18-crown-6)Rb- and Cs+(C222)Cs-, in which linear chains of Rb- and Cs-, respectively, are found.9,11 Lowdimensional nanoscale materials are currently receiving considerable experimental and theoretical attention, and the existence of these one-dimensional (1D) systems of alkali metal anions raises pertinent questions about their properties and electronic structures. In the present work, we report the synthesis, crystal structure, and properties of a new compound, Li+(cryptand[2.1.1])Cs(abbreviated Li+(C211)Cs-), which is the second ceside and the third alkalide that has a linear zigzag chain of alkali metal * Authors to whom correspondence should be addressed. E-mail: [email protected]; [email protected]. † Michigan State University. ‡ San Francisco State University.

anions. Although the structure and some properties of Cs+(cryptand[2.2.2])Cs- (abbreviated Cs+(C222)Cs-), have been previously reported,11,12 we include here a number of new measurements and interpretations. From a materials science perspective, these two cesides are intriguing because they contain the largest monatomic anion, Cs-, which should be highly polarizable, and a complexed cation that balances the negative charge. The Herzfeld criterion for metallization relates the static polarizability, R, to the molar volume, V, through 4πNR0/3V ) R/V and states that if this ratio is greater than one, then the substance will be metallic.13,14 Part of the impetus for synthesizing Li+(C211)Cs- was to probe the Herzfeld criterion with a novel material. To this end, the temperature-dependent powder resistivities of both cesides were measured. The primary diagnostic for alkali metal anions is the NMR chemical shift.12 The closed-shell ns2 electron configuration diamagnetically shields the nucleus and results in a substantial upfield resonance. For example, the chemical shift of Na- is at -60 to -65 ppm relative to Na+(aq), both in solution and in the solid state.15 The chemical shifts of the larger anions, Rb-, Cs-, and to a lesser extent K-, are more sensitive measures of

10.1021/jp061200l CCC: $33.50 © 2006 American Chemical Society Published on Web 06/02/2006

12294 J. Phys. Chem. B, Vol. 110, No. 25, 2006 environmental effects, such as the solvent dieletric constant, concentration in solution, temperature, and the local surroundings within a crystal. Nevertheless their large negative chemical shifts (relative to the cation at infinite dilution) permit identification of the alkali anion. The proximity of alkali anions in the solids that have dimers and chains results in large Ramsey shifts16 of their NMR resonances relative to M- in dilute solutions or in structures that have isolated anions.10,12,17 Note that the complexed cations also have substantial Ramsey NMR chemical shifts, but these are much less sensitive to their environment because they are sequestered inside complexant molecules. The optical absorbance spectrum of Cs+(C222)Cs- thin films was first reported in 197918 and was unusual for an alkalide because at least three peaks were observed. In the more typical case where an alkalide is in a nearly isotropic environment, a single peak ascribed to the ns-to-np transition is observed. At the time, the structure of Cs+(C222)Cs- had not been determined. With the current insight gained from the crystal structure, it is known that the 1D chain provides an anisotropic environment about Cs- such that the degeneracy of the p-orbitals should be lifted. From this perspective, multiple absorptions might be expected. In both Li+(C211)Cs- and Cs+(C222)Cs-, the question of optical anisotropy may be answered by single-crystal reflectance spectroscopy using plane-polarized light as was done for Na+(C222)Na-.19 The electronic structure of potasside dimers, K22-, was studied by ab initio calculations.20 A valence bond interpretation accounted for the stability of the anion dimers by placing two of the electrons in a σ-bond between each K- and the other two electrons in localized nonbonding sp-hybrid orbitals that point away from the nuclei along the internuclear axis toward the point positive charges required for stability of the dimer. Although this view is oversimplified by neglect of the effect of the position of the complexed cations, the model provides a convenient physical picture to understand the stability of K22despite the Coulomb repulsion between anions. In the present work, the preliminary results of multiconfiguration selfconsistent field (MCSCF) calculations for Cs- trimers, Cs33-, are reported. To derive meaningful comparisons between Li+(C211)Csand Cs+(C222)Cs-, fresh samples of the latter compound were prepared and characterized. The full suite of experimental methods included X-ray diffraction, single-crystal reflectance and thin film absorbance spectroscopy, differential scanning calorimetry (DSC), electron paramagnetic resonance (EPR), 133Cs and 7Li magic angle spinning nuclear magnetic resonance (MAS-NMR), magnetic susceptibility, and direct current (DC) conductivity measurements. These experiments in conjunction with ab initio computations permit insight into the intrinsic properties and the nature of defects in ceside chains. Experimental Methods Synthesis. Cryptand[2.1.1] and cryptand[2.2.2] were purchased from Aldrich and purified by sublimation. Lithium was purchased from AESAR and used as received. Cesium metal was a gift from the Dow Chemical Co. The synthesis of Li+(C211)Cs- proceeded as follows: 1.4 mmol of lithium and C211 were placed on one side of a twochambered K-cell,4,21 and 1.4 mmol of cesium was placed in a sidearm on the opposite side. The cesium was distilled to form a mirror, and the sidearm was sealed off from the K-cell. Methylamine was distilled onto the lithium metal and C211 at -70 °C. After all lithium metal and C211 had been dissolved,

Ichimura et al. the dark blue solution was poured through a frit onto the cesium mirror. The methylamine was removed under vacuum overnight and then dimethyl ether was added and pumped away to remove lingering methylamine. Approximately 25 mL of fresh dimethyl ether was added and the solid was dissolved at -78 °C. To form a saturated solution of the ceside, diethyl ether was added and then some dimethyl ether was removed. Crystals up to 2 cm in length were grown by slowly removing the majority of the primary solvent over a period of 2 days at -78 °C. Cs+(C222)Cs- was prepared according to a previously published method.21 The crystal structure was determined with a Bruker CCD diffractometer and low-temperature accessory. The Li+(C211)Cs- crystals were kept at or near liquid nitrogen temperature under vacuum or a dry nitrogen atmosphere at all times during transfer to sample holders. Fresh red-gold, gold, and black crystals were placed in a mortar precooled to 77 K in a nitrogen-purged glovebag. At this stage, the crystals were hand-sorted by color. The majority of the crystals were red-gold while some gold crystals and a few small black crystals were obtained. The red-gold and gold crystals were crushed with a cold pestle and placed in the appropriate sample holders at 77 K for solid-state MAS-NMR, EPR, DSC, magnetic susceptibility, and DC conductivity measurements. Magic angle spinning NMR, DSC, and conductivity samples were run immediately, but the powder EPR samples were sealed off under vacuum and stored at 77 K until measured. Magic Angle Spinning NMR. Crushed red-gold crystals and intact black crystals were loaded into separate ZrO2 rotors and placed into liquid nitrogen. The NMR cryostat was cooled to -70 °C, and the samples were transferred rapidly to the probe. The sample spin rate was changed to ensure correct identification of the isotropic chemical shift. A Varian VXR 400S spectrometer was used to acquire 7Li and 133Cs MAS-NMR spectra. Differential Scanning Calorimetry. A small amount of crushed crystals was placed in an aluminum pan, covered with an aluminum lid, and loaded into the sample press. The sample holders and press were kept in contact with liquid nitrogen in a glovebag to prevent thermal decomposition of the alkalide. The low-temperature accessory of the Shimadzu DSC-50 calorimeter was used to measure thermal data from -80 to 50 °C. Two-Probe Direct Current Conductivity and Alternating Current Impedance Spectroscopy. The two-probe conductivity cell has two gold-coated electrodes, one fixed in position and one that is spring-loaded. The sample diameter is 2 mm, and the final height was 3 mm. Before the ceside powder was packed into the cell, the tips of the gold electrodes were coated with a small quantity of potassium metal to ensure good electrical contact. This has the added benefit that electrode effects, such as the formation of a Schottky barrier at the probe-sample junction, are minimized.22 For the two-probe impedance measurements, the frequency was stepped logarithmically between 5 Hz and 11.7 MHz at voltages between 50 mV and 1 V. A Hewlett-Packard 4192A low-frequency impedance analyzer interfaced to the LabView program was utilized for impedance measurements. For DC measurements, the conductivity cell was interfaced to a Keithley electrometer via LabView, and the current was measured with a (0.1 V bias as a function of temperature. Electron Paramagnetic Resonance. Individual gold, redgold, and black Li+(C211)Cs- single crystals were placed in separate 4 mm o.d. quartz tubes and sealed off under vacuum. The preparation of powdered ceside EPR samples was described above. Spectra were recorded with a Bruker ESP300E spec-

One-Dimensional Zigzag Chains of Cs-

J. Phys. Chem. B, Vol. 110, No. 25, 2006 12295

TABLE 1: Crystallographic Data for Li+(Cryptand[2.1.1])Cscell parameters

temperature space group Z number of unique reflections: number of parameters: R for reflections with I > 2σ R for all reflections

a ) 23.748(15) b ) 13.434(8) c ) 8.212(12) β ) 75.34(9)° V ) 2534.6(43) Å3 173 K monoclinic, C2/m 4 1752 106 0.120 0.167

trometer, and the temperature was controlled with an Oxford Instruments 900E liquid helium cryostat or a Bruker VT2000 nitrogen gas flow cryostat. Optical Spectra of Crystals and Thin Films. Single crystals of freshly prepared Li+(C211)Cs- and Cs+(C222)Cs- were transferred into a quartz optical cell (1 cm path length) that was attached directly to the synthesis vessel (K-cell). The optical cell was sealed off under vacuum at -78 °C thus ensuring that the surface of the crystals would remain pristine. The optical cell was kept cold during specular reflectance measurements by placing the optical cell in direct contact with a cold copper block (∼-70 to -40 °C) that was mounted on a 3D translational stage and cooling with a stream of cold nitrogen gas in a nitrogen-purged glovebag. The Cassegrainian optics for reflectance studies were also kept in the glovebag, and light was passed to a Guided Wave model 260 spectrophotometer via optical fibers. For these studies, crystals were chosen based on their overall reflecting ability to increase the signal-to-noise ratio. Because of the geometry of the crystals only a single face could be studied, and the Miller indices were unknown. In addition, because of instrument limitations, the crystal was rotated rather than the plane of polarization. Ceside thin films were prepared by flash evaporation of a dimethyl ether solution, in which a few small, freshly prepared red-gold crystals were dissolved in approximately 1 mL of solvent. The solution was poured into a 1 cm path length quartz cell, and the temperature was raised to about -35 °C. The larger bulb of the optical cell was then immersed in liquid nitrogen, which caused solvent evaporation and thin film formation in a matter of seconds. This optical cell was then placed in a stream of cold N2 gas in a home-built quartz dewar, and absorbance spectra were recorded. Results and Discussion Li+(Cyptand[2.1.1]Cs- Crystal Structure and Diagrams. Two types of crystals were formed in one preparation, elongated rectangular gold and red-gold crystals and small blocky black crystals. The structure of the gold crystals was determined. The gold and red crystals differ mainly in the number of defect centers, vide infra. The crystal system is monoclinic with space group C2/m (No. 12). The cell parameters are listed in Table 1, and the CIF file is given in the Supporting Information. The striking feature of the crystal structure, Figure 1A, is that like charges get together to form zigzag chains of cesium anions with a center-to-center distance of 6.0 Å. This is remarkable because the van der Waals radius of Cs- in crystal structures that have isolated ceside anions, such as Cs+(18crown-6)2Cs-,9 in which each anion is surrounded by complexed cations, is 3.5 Å. Therefore, in Li+(C211)Cs-, adjacent cesium anions overlap by approximately 1.0 Å. The angle between three adjacent Cs- ions is 86.7°. This is the third alkalide crystal structure to have linear chains of M-. Cs+(C222)Cs- has a

Figure 1. Ion and atom packing in Li+(C211)Cs-: (A) perpendicular to the chains; (B) a view down the chains. Cs, Li, N, O, and C atoms are purple, green, blue, red, and gray, respectively.

similar Cs- chain structure in which the distance between pairs of anions is 6.38 Å and the angle between three neighbors is 118.0°.11 In Rb+(18-crown-6)Rb-, the Rb--to-Rb- distance is 5.13 Å, the same as that in rubidide pairs in Rb+(C222)Rb-.9 From the standpoint of cell type and parameters, the ceside crystal structure resembles that of Li+(cryptand[2.2.1])e-.23 In this electride, the trapped electrons form 1D antiferromagnetically coupled chains, and the arrangement of complexed cations relative to the electrons strongly resembles the cation-anion placement in the ceside, Figure 1B. The encapsulated Li+ fragment of the ceside salt is typical of other alkalide salts and the electride that contains the Li+(cryptand[2.1.1]) cation. Oxygen and tertiary nitrogen atoms of the complexant point inward stabilizing Li+ in a hexacoordinate arrangement. Stabilizing charge balance for the anionic chain is provided by interleaved complexed cations that form their own 1D zigzag arrays. The alternating arrays of positive and negative charge likely lead to the brittleness of the crystals, as they tend to fragment lengthwise when a mild force is applied. The structure of the black crystals was not solved. However, they strongly resemble Li+(cryptand[2.1.1])e- in appearance. The black color is generally associated with a high concentration of defect electrons. The solid-state NMR and EPR spectra of this form of “doped” ceside provide useful information about the nature of chain-end and electron defects in alkali metal anion chains. Magic Angle Spinning NMR. The 133Cs MAS-NMR spectrum of crushed gold-red Li+(C211)Cs- crystals yields a temperature-independent 133Cs chemical shift of +72.6 ppm (Figure 2a). The chemical shift of Cs- in crystalline Cs+(C222)Cs- is even larger at +127.4 ppm (Figure 2b). In a previous publication,12 the peak at 127.4 ppm was interpreted as an “exclusive” complex of Cs+ with cryptand[2.2.2], that is, one in which Cs+ was complexed by C222 in the manner of a crown ether rather than a cryptand. In this work, optical spectra in combination with the NMR data show conclusively that this peak is due to a large downfield shift of Cs- relative to Cs-(g). Relative to the isolated Cs- in Cs+(18-crown-6)2Cs- (-213 ppm),12 these values represent overall paramagnetic shifts of +285.6 and +340.4 ppm for the Cs- anion in Li+(C211)Csand Cs+(C222)Cs-, respectively. The inclusive complex of the cesium cation, Cs+(cryptand[2.2.2]), experiences a large Ramsey shift of +238.3 ppm, which is typical for Cs+ sequestered inside this complexant.24 In previous work,25 it was found that the quadrupole coupling of the cation is large so that only the central transition shows, while that of the cesium anion is small and results in all transitions being involved. This accounts for the larger magnitude and anisotropy of the NMR spectrum of the anion. The lithium cation complexed by cryptand[2.1.1] in gold crystals of Li+(C211)Cs- has a small temperature-independent

12296 J. Phys. Chem. B, Vol. 110, No. 25, 2006

Ichimura et al.

Figure 3. Differential scanning calorimetry trace of crushed gold crystals of Li+(C211)Cs-.

Figure 2. 133Cs MAS-NMR spectra of (a) Li+(C211)Cs- and (b) Cs+(C222)Cs-. A single dot signifies the assignment to Cs-. Two dots identify the peak of complexed Cs+. Others are spinning sidebands.

chemical shift of -2.1 ppm relative to the value of 7Li+ at infinite dilution. This behavior is typical for the Li+(C211) complex in other alkalides, such as Li+(C211)Na- and Li+(C211) in solutions. Quite the opposite behavior was observed for the strongly temperature-dependent 7Li chemical shift of the black crystals. This chemical shift, extrapolated to high temperatures (1/T ) 0) yields σ(∞) ) -0.3 ( 3 ppm, which is comparable to that of Li+(cryptand[2.1.1]) complexes in solution. At 233 K, the 7Li chemical shift for the black crystals was +20 ppm relative to Li+(aq,∞), which is approximately one-third of the value for Li+ in the electride, Li+(C211)e-.23 The electride has one electron per complexed cation and has a 7Li chemical shift of 57 ppm at 233 K. Since Li+(C211)Cshas one-third of this value, it suggests that there is on average one-third of an electron per Li+, a very high concentration of electrons in an alkalide. These small blocky black crystals of Li+(C211)Cs- did not exhibit a 133Cs resonance, probably due to paramagnetic broadening by the unpaired electrons. Such paramagnetic broadening into the baseline has been seen for cesium in a zeolite.26 Magnetic susceptibility measurements on crushed crystals show that both cesides have low paramagnetic defect concentrations. The susceptibility accounted for 3, and the band structure of the bulk. Ab initio and density functional theory calculations that address these issues are in progress and will be reported in a subsequent paper. Conclusions Charge transport within alkalides and electrides usually involves defect centers in the crystal. For example, K+(C222)eis nearly metallic and has a 2D pore structure with anisotropic conductivity.32 The conductivity in this solid most likely arises by a hole percolation mechanism since the temperature dependence is well fit by a random 2D hopping model. This points to an electron vacancy as one of the defects in this electride. The defects in Li+(C211)Cs- are primarily localized electrons at chain ends with a very small percentage of defect electrons in anion vacancies. Although the EPR spectrum of Cs+(C222)Cssingle crystals was not measured, based on the crystal structure it is reasonable to suppose that chain ends are also the main defects in this solid. Chain ends imply a ceside vacancy otherwise a diamagnetic state would result. The temperature dependence of the conductivity suggests that charge transport involves the thermal population of a conduction band rather than a hopping mechanism. In addition, the optical reflectance spectra indicate a plasmalike resonance at long wavelengths. From these data and an understanding of the defects in these solids, it is possible to conclude that the semiconducting behavior is an intrinsic property of the 1D ceside chains. Furthermore, because of the anisotropy in the single-crystal reflectance measurements, it is quite likely that charge transport would be anisotropic. According to our crushed crystal DC conductivity studies, these solids are on the R/V < 1 side of the Herzfeld criterion. Thus, they are nonmetals. However, to pin down the mechanism, verify transport directionality, and provide a better test of the Herzfeld inequality, single-crystal conductivity studies on high-quality gold crystals would be necessary. The two cesides, Li+(C211)Cs- and Cs+(C222)Cs-, differ in the angle and amount of overlap between adjacent centers. Nuclear magnetic resonance samples the local environment of the nuclei. It is hypothesized that the larger Ramsey shift of Cs+(C222)Cs- compared to Li+(C211)Cs- correlates with the

Ichimura et al. bond angles, which are 118° and 86.7°, respectively. In this model, Cs+(C222)Cs- has greater p-character that deshields the 133Cs nucleus to a larger extent than in Li+(C211)Cs-. The distance between Cs- may correlate with the conductivity, since Cs+(C222)Cs- powders conduct less than Li+(C211)Cs- at a given temperature. A larger distance decreases the overlap between p-orbitals on adjacent anions, and so more thermal energy is required to activate charge transport. An electronic structure model that is consistent with the experimental data and supported by ab initio calculations has sp2-hybridized Cs- within the chain and sp-hybridized chain ends. As a result, Li+(C211)Cs- and Cs+(C222)Cs- have dramatically different properties from alkalide monomers (isolated M-) and dimers, such as Na22-. This study has revealed surprising details about 1D alkalide chains, but much experimental and theoretical work is needed to fully understand these materials. Acknowledgment. This work was supported in part by the U. S. National Science Foundation under Grant No. DMR 9988881. The American Chemical Society Petroleum Research Fund Grant No. 38397-GB5 and Research Corporation Cottrell Science Award No. CC5646 also supported this work (A.S.I.). Supporting Information Available: Crystallographic data and EPR spectra of single crystals of Li+(C211)Cs-. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Dye, J. L.; Ceraso, J. M.; Lok, M. T.; Barnett, B. L.; Tehan, F. J. J. Am. Chem. Soc. 1974, 96, 608-609. (2) Tehan, F. J.; Barnett, B. L.; Dye, J. L. J. Am. Chem. Soc. 1974, 96, 7203-7208. (3) Dye, J. L. Sci. Am. 1977, 237, 92-105. (4) Dye, J. L. Prog. Inorg. Chem. 1984, 32, 327-441. (5) Dye, J. L. Chemtracts: Inorg. Chem. 1993, 5, 243-270. (6) Huang, R. H.; Huang, S. Z.; Dye, J. L. J. Coord. Chem. 1998, 46, 13-31. (7) Kim, J.; Ichimura, A. S.; Huang, R. H.; Redko, M.; Phillips, R. C.; Jackson, J. E.; Dye, J. L. J. Am. Chem. Soc. 1999, 121, 10666-10667. (8) Wagner, M. J.; Dye, J. L. In Molecular Recognition: Receptors for Cationic Guests, 1st ed.; Gokel, G. W., Ed.; Pergamon Press: Oxford, U. K., 1996; Vol. 1, pp 477-510. (9) Huang, R. H.; Ward, D. L.; Dye, J. L. J. Am. Chem. Soc. 1989, 111, 5707-5708. (10) Redko, M. Y.; Huang, R. H.; Jackson, J. E.; Harrison, J. F.; Dye, J. L. J. Am. Chem. Soc. 2003, 125, 2259-2263. (11) Huang, R. H.; Ward, D. L.; Kuchenmeister, M. E.; Dye, J. L. J. Am. Chem. Soc. 1987, 109, 5561-5563. (12) Dawes, S. B.; Ellaboudy, A. S.; Dye, J. L. J. Am. Chem. Soc. 1987, 109, 3508-3513. (13) Herzfeld, K. F. Phys. ReV. 1927, 29, 701-705. (14) Edwards, P. P.; Sienko, M. J. Int. ReV. Phys. Chem. 1983, 3, 83137. (15) Ellaboudy, A.; Dye, J. L. J. Magn. Reson. 1986, 66, 491-502. (16) Ramsey, N. F. Phys. ReV. 1950, 78, 699-703. (17) Kim, J.; Eglin, J. L.; Ellaboudy, A. S.; McMills, L. E. H.; Huang, S.; Dye, J. L. J. Phys. Chem. 1996, 100, 2885-2891. (18) DaGue, M. G.; Landers, J. S.; Lewis, H. L.; Dye, J. L. Chem. Phys. Lett. 1979, 66, 169-172. (19) Hendrickson, J. E.; Kuo, C. T.; Xie, Q.; Pratt, W. P., Jr.; Dye, J. L. J. Phys. Chem. 1996, 100, 3395-3401. (20) Tientega, F.; Dye, J. L.; Harrison, J. F. J. Am. Chem. Soc. 1991, 113, 3206-3208. (21) Dye, J. L. J. Phys. Chem. 1984, 88, 3842-3846. (22) Moeggenborg, K. J.; Papaioannou, J.; Dye, J. L. Chem. Mater. 1991, 3, 514-520. (23) Huang, R. H.; Wagner, M. J.; Gilbert, D. J.; Reidy-Cedergren, K. A.; Ward, D. L.; Faber, M. K.; Dye, J. L. J. Am. Chem. Soc. 1997, 119, 3765-3772. (24) Dye, J. L.; DeBacker, M. G. Annu. ReV. Phys. Chem. 1987, 38, 271-301.

One-Dimensional Zigzag Chains of Cs(25) Kim, J. Solid State Alkali Metal NMR Spectra and Magnetic Susceptibilities of Alkalides, Electrides and Their Related Compounds. Ph.D. Dissertation, Michigan State University, East Lansing MI, 1989. (26) Wernette, D. P.; Ichimura, A. S.; Urbin, S. A.; Dye, J. L. Chem. Mater. 2003, 15, 1441-1448. (27) Dye, J. L.; Yemen, M. R.; DaGue, M. G.; Lehn, J.-M. J. Chem. Phys. 1978, 68, 1665-1670. (28) Dye, J. L.; DaGue, M. G.; Yemen, M. R.; Landers, J. S.; Lewis, H. L. J. Phys. Chem. 1980, 84, 1096-1103. (29) Redko, M. Y.; Jackson, J. E.; Huang, R. H.; Dye, J. L. J. Am. Chem. Soc. 2005, 127, 12416-12422.

J. Phys. Chem. B, Vol. 110, No. 25, 2006 12301 (30) Hendrickson, J. E.; Pratt, W. P., Jr.; Phillips, R. C.; Dye, J. L. J. Phys. Chem. B 1998, 102, 3917-3926. (31) Ichimura, A. S.; Wagner, M. J.; Dye, J. L. J. Phys. Chem. B 2002, 106, 11196-11202. (32) Shin, D.-H.; Ellaboudy, A. S.; Dye, J. L.; DeBacker, M. G. J. Phys. Chem. 1991, 95, 7085-7089. (33) (a) Werner, H.-J.; Knowles, P. J. J. Chem. Phys. 1985, 82, 5053. (b) Knowles, P. J.; Werner, H.-J. Chem. Phys. Lett. 1985, 115, 259. (34) (a) For K: Bergner, A.; Dolg, M.; Kuechle, W.; Stoll, H.; Preuss, H. Mol. Phys. 1993, 80, 1431. (b) For Ca: Kaupp, M.; Schleyer, P. v. R.; Stoll, H.; Preuss, H. J. Chem. Phys. 1991, 94, 1360. (c) For Rf-Db: Dolg, M.; Stoll, H.; Preuss, H.; Pitzer, R. M. J. Phys. Chem. 1993, 97, 5852.